European Journal of Vascular and Endovascular Surgery

Transcription

1 European Journal of Vascular and Endovascular Surgery 43 (2012) 139e145 Contents lists available at SciVerse ScienceDirect European Journal of Vascular and Endovascular Surgery journal homepage: The War Against Error: A 15 Year Experience of Completion Angioscopy Following Carotid Endarterectomy R. Sharpe, R.D. Sayers, M.J. McCarthy, M. Dennis, N.J.M. London, A. Nasim, M.J. Bown, A.R. Naylor * The Department of Vascular Surgery, Clinical Sciences Building, Leicester Royal Infirmary, Leicester, UK WHAT THIS PAPER ADDS This study has shown that some degree of inadvertent technical error (the principle cause of intra-operative stroke during carotid endarterectomy) occurs in about 10% of procedures. While the prevalence of large intimal flaps decreased with time, the prevalence of retained luminal thrombus prior to restoration of flow did not, a reminder to even the most experienced surgeon that he/she can still be responsible for inadvertent technical error. article info abstract Article history: Received 1 August 2011 Accepted 8 September 2011 Available online 5 October 2011 Keywords: Carotid endarterectomy Quality control Angioscopy Stroke Background: A policy of intra-operative transcranial Doppler (TCD) and completion angioscopy was previously associated with virtual abolition of intra-operative stroke (apparent upon recovery from anaesthesia) following carotid endarterectomy (CEA). The aims of this study were to determine whether the prevalence of technical error has diminished with experience and whether our monitoring/quality control policy was still associated with low rates of intra-operative stroke 20 years after its introduction. Methods: Retrospective review of four consecutive cohorts of 400 patients undergoing CEA between October 1995 and March 2010 (1600 CEAS in total). Results: One hundred four patients (7%) had thrombus removed following angioscopy and prior to flow restoration, while 31 (2.1%) underwent repair of a distal intimal flap. The prevalence of intimal flaps diminished from 4.9% in the first 400 patients to 0.8% in the last 400 patients (p ¼ 0.006). By contrast, the prevalence of retained thrombus did not decline with experience (8.5%, 3.7%, 10.3% and 5.4% for the four consecutive periods). Intra-operative TCD and completion angioscopy was, however, associated with extremely low rates of intra-operative stroke (0.25%, 0.25%, 0.5% and 0.25% during the four study periods). Conclusion: Most intra-operative strokes probably follow embolisation of thrombus following restoration of flow. This can be prevented by angioscopy which has the advantage of being performed prior to flow restoration. Increasing experience was associated with a decline in the detection of intimal flaps, but not in the prevalence of retained thrombus. Even the most experienced of surgeons can still be responsible for inadvertent technical error. Ó 2011 European Society for Vascular Surgery. Published by Elsevier Ltd. All rights reserved. Introduction Carotid endarterectomy (CEA) is a proven intervention for preventing stroke in selected patients with atherosclerotic carotid disease. 1,2 However, while the magnitude of benefit varies within patient subgroups (eg greater benefit in males and those * Corresponding author. Tel.: þ ; fax: þ address: ross.naylor@uhl-tr.nhs.uk (A.R. Naylor). undergoing surgery within 14 days of the index event 3 ), the overall benefit is inextricably linked to the procedural risk. This means that the higher the initial risk, the lower the number of strokes prevented in the long term. In a series of themed projects, this unit has shown that most intra-operative strokes (ie apparent upon recovery from anaesthesia) follow inadvertent technical error and that they can be virtually abolished by an intra-operative monitoring programme comprising transcranial Doppler ultrasound (TCD) and completion angioscopy. 4e6 By contrast, post-operative thromboembolic strokes /$ e see front matter Ó 2011 European Society for Vascular Surgery. Published by Elsevier Ltd. All rights reserved. doi: /j.ejvs

2 140 R. Sharpe et al. / European Journal of Vascular and Endovascular Surgery 43 (2012) 139e145 were found to be unrelated to technical error 6 (ie they could never be prevented by angioscopy), but they seemed to be associated with an increased susceptibility of the patient s platelets to ADP, 7 leading to the development of novel antiplatelet strategies in order to prevent them. 8,9 Following a pilot study and subsequent audit, 4,5 this Unit continued to perform completion angioscopy in as many patients as possible undergoing CEA in order to identify and correct technical error. The aims of the study were to determine whether; (i) the prevalence of technical error detected using completion angioscopy had diminished with increasing experience (ie was it still necessary to perform angioscopy) and (ii) was the extremely low rate of intra-operative stroke maintained through the ensuing 15 year period? Materials and Methods A retrospective audit of prospectively acquired data in 1600 consecutive patients undergoing CEA at the Leicester Royal Infirmary between 1st October 1995 and 10th March 2010 was performed with the aim of determining whether the prevalence of technical error was reduced through the continued application of a quality control programme. The Leicestershire, Northamptonshire and Rutland Research Ethics Committee advised that this study did not fall under the remit of the NHS Research Ethics Committee as it was audit/service evaluation. Carotid endarterectomy This was a consecutive series of 1600 patients undergoing a primary CEA for atherosclerotic disease, including 48 who ultimately underwent interposition vein grafts due to marked coiling or excessive thinning of the residual wall after endarterectomy. The basic technique of CEA (general anaesthesia, systemic heparinisation, loupe magnification, routine shunting, routine patching and distal intimal tacking) has remained unchanged since All patients received 75e150 mg Aspirin throughout the peri-operative period. TCD monitoring was commenced after induction of anaesthesia using a fixed 2 MHz head probe, which was protected by a semi-circular headguard. The surgeon and anaesthetist tried to ensure that mean blood flow velocity in the middle cerebral artery (MCAV) was >15 cm/s at all times. This threshold of 15 cm/s was chosen because Halsey has shown this to correlate with loss of cerebral electrical activity. 10 If the mean MCAV was <15 cm/s following shunt insertion, the shunt was repositioned to exclude abutment against the distal ICA lumen. If the MCAV remained <15 cm/s, blood pressure was therapeutically elevated by the anaesthetist. Immediately prior to completion of the patch closure, a 5 mm space was retained adjacent to the orifice of the external carotid artery. The shunt was then removed and all vessels back vented and irrigated with heparinised saline. The lumen of the endarterectomy zone (Fig. 1) was then inspected with a flexible hysteroscope (Olympus ,). Our policy was to repair all intimal flaps >3 mm and to aspirate any residual thrombi (irrespective of size) from the lumen. Following angioscopy, the patch was closed and flow was restored first up the external carotid artery (with a finger compressing the proximal ICA) and then up the ICA. Post-operative monitoring Between October 1st 1995 and 31st July 2006, patients were transferred through to the recovery area of the operating theatre for a 3 h period of post-operative TCD monitoring in order to identify patients with increasing rates of embolisation. 11 Previous Figure 1. The carotid patch has almost been completed and the shunt has been removed. The vessels have been irrigated with heparinised saline and the lumen is about to be inspected with a hysteroscope prior to restoration of flow. work has shown that patients with increasing rates of embolisation following CEA are at high risk of progressing onto carotid thrombosis 12e19 and that progression towards stroke can be prevented by incremental dose Dextran therapy. 20,21 However, following the manufacturer s warning that it was ceasing manufacture of Dextran after August 2006, all CEA patients subsequently received 75 mg Clopidogrel the night before surgery (in addition to their aspirin) in order to prevent post-operative thromboembolic stroke. 9 The only other significant change in post-operative care was the introduction of written guidelines for the treatment of post-operative hypertension in July Definition of post-operative stroke Any new neurological deficit lasting more than 24 h and which occurred within 30 days of CEA was defined as a post-operative stroke. Intra-operative stroke was defined as having occurred where the deficit was apparent upon recovery from anaesthesia. A post-operative stroke was deemed to have occurred if a neurological deficit developed after an otherwise uneventful recovery from anaesthesia. Wherever possible, the most likely cause of the stroke was determined by autopsy and/or CT scanning, Duplex imaging of the extracranial carotid circulation and transcranial Doppler insonation of flow patterns in the circle of Willis. For the purposes of this study, a post-operative stroke was attributed to MCA embolisation from the endarterectomy zone if there was a patent ICA on Duplex imaging, no evidence of a recent myocardial infarction, but a focal area of infarction in the ipsilateral MCA territory irrespective of whether TCD had demonstrated post-operative embolisation. Stroke due to post-operative carotid thrombosis (POCT) was diagnosed where re-exploration and/or Duplex imaging suggested carotid thrombosis. A diagnosis of hyperperfusion syndrome was made in patients who presented with a new neurological deficit, preceded by either worsening headaches and/or seizures in the absence of embolisation on TCD or haemorrhage on CT scanning. An unknown aetiology was documented in patients who had no evidence of a cardioembolic source of embolism, and who had also undergone a normal CT scan, extracranial Duplex scan and TCD study. Results Between 1st October 1995 and 10th March 2010, 1600 consecutive patients underwent CEA in the Vascular Unit at Leicester Royal Infirmary. In addition to analysing data for the 15-year period, four

3 R. Sharpe et al. / European Journal of Vascular and Endovascular Surgery 43 (2012) 139e Table 1 Gender and mode of presentation in 1600 patients undergoing CEA between 1995 and e e e e Total series (n ¼ 1600) Gender Male 263 (66%) 279 (70%) 292 (73%) 273 (68%) 1107 (69%) Female 137 (34%) 121 (30%) 108 (30%) 127 (32%) 493 (31%) Presentation Stroke 118 (30%) 103 (26%) 96 (24%) 79 (20%) 396 (25%) TIA/amaurosis 216 (54%) 257 (64%) 233 (58%) 254 (64%) 960 (60%) Asymptomatic 66 (16%) 33 (8%) 66 (17%) 67 (17%) 232 (14%) Uncertain 7 (2%) 5 (1%) 12 (1%) consecutive groups of 400 patients were identified in order to ascertain temporal changes in angioscopy findings ( e ; e ; e and e ). Basic patient demographics are summarised in Table 1. Overall, the majority were male (69%), 396 (25%) presented with a stroke, while 960 (60%) presented with either a transient ischaemic attack or amaurosis fugax. Only 232 patients (15%) were neurologically asymptomatic prior to surgery. Completion angioscopy was undertaken in 1483 patients (93%). There were a number of reasons why 117 patients (7%) did not undergo angioscopic assessment including; simultaneous CEAs and only one angioscope available (partially resolved by purchasing a second hysteroscope in 2002), occasional problems with the sterilisation process and one/both angioscopes malfunctioning or being broken. Out of the 1483 patients undergoing angioscopy (Table 2), 1343 (91%) were normal (Fig. 2). However, retained luminal thrombus (Figs. 3 and 4) was aspirated in 104 patients (7%) prior to restoration of flow. Thirty-one patients (2.1%) underwent reexploration to repair a distal intimal flap (Fig. 5). Luminal thrombus was observed throughout the endarterectomy zone, but was seldom adherent to the patch. Two sources of thrombus were identified in this study. Fig. 3 shows a tubular peri-shunt thrombus that formed around the upper reaches of the shunt during the endarterectomy period. When the shunt was removed and the vessel flushed, the thrombus dropped into the endarterectomy zone where it became caught in the distal clamp. It only became apparent to the surgeon following angioscopy. Fig. 4a shows thrombus adherent to the proximal endarterectomy zone. Following aspiration, the source of the thrombus was seen to be bleeding from the transected vasa vasorum. Thrombi that were adherent to the endarterectomy zone were remarkably resistant to flushing and blind irrigation with heparinised saline prior to restoration of flow. Note that there has not been a consistent reduction in the prevalence of retained thrombus throughout the entire study period (still 5% in the final period of study), but there was a significant reduction in the prevalence of large intimal flaps requiring repair (4.9% in the first period compared with 0.8% in the final period (p ¼ 0.006)). This reduction in the prevalence of distal intimal flaps with time was probably because surgeons paid Figure 2. Normal angioscopy of the distal endarterectomy zone and internal carotid artery. greater attention to avoiding situations where flaps might develop (better feathering/trimming of the distal intimal step, better placement of tacking sutures). Table 3 details rates of intra-operative stroke, 30-day death and 30-day death/stroke for each time period and also for the entire series. As can be seen, the 30-day rates of death/stroke were relatively constant for the first three time periods (averaging 2.5%), but then reduced to 0.75% for the final 400 patients. The rate of intraoperative stroke has remained extremely low across all four time periods (0.3%). The main role for completion angioscopy is the identification and correction of technical error before it can be responsible for causing an intra-operative stroke. Accordingly, Table 4 details the causes of stroke/death in the 1600 patients, with particular emphasis on the most likely aetiology during the four time periods. Note that 5/1600 patients (0.3%) recovered from anaesthesia with a new neurological deficit that persisted for >24 h. Two of these strokes were strokes were fatal, two were disabling and one was non-disabling. One patient (in the first period of study) underwent a normal angioscopy following vein bypass, but then proceeded to thrombose the bypass on two occasions before the operation was completed. On-table angiography and repeat angioscopies revealed no evidence of any technical error and we assume that he was one of the very rare cases with a patient mediated prothrombotic disorder. Interestingly, the diagnosis of on-table thrombosis was only possible because of the availability of TCD monitoring. Ultimately, with intravenous Dextran and Heparin to control the embolisation, it was possible to complete the procedure and he recovered with a non-disabling stroke. Table 2 Results of completion angioscopy following carotid endarterectomy e e e e Total 1995e2010 Number of CEAs Angioscopy performed 386 (96.5%) 349 (87.25%) 360 (90.0%) 388 (97.0%) 1483 (92.7%) Normal angioscopy 334/386 (86.5%) 330/349 (94.5%) 315/360 (87.5%) 364/388 (93.8%) 1343/1483 (90.5%) Luminal thrombus 33/386 (8.5%) 13/349 (3.7%) 37/360 (10.3%) 21/388 (5.4%) 104/1483 (7.0%) Intimal flap 14/386 (4.9%) 6/349 (1.7%) 8/360 (2.2%) 3/388 (0.8%) 31/1483 (2.1%)

4 142 R. Sharpe et al. / European Journal of Vascular and Endovascular Surgery 43 (2012) 139e145 Figure 3. Angioscopy of a patient who suffered a peri-shunt thrombosis. Following shunt removal and back bleeding, the thrombus fell down into the endarterectomy zone where it was (unknowingly) held within the distal internal carotid artery clamp. In the second study period, one patient suffered a disabling stroke secondary to acute MCA occlusion immediately following flow restoration. This was diagnosed using TCD. Unfortunately, this occurred in one of the 117 patients who did not undergo completion angioscopy (broken). If he had undergone angioscopic assessment, it is likely that the residual luminal thrombus (which embolised following flow restoration) would have been detected and removed. In the third period of study two fatal intra-operative strokes occurred. The first followed acute MCA occlusion (diagnosed using TCD) immediately following shunt insertion. It is assumed that the shunt inadvertently ploughed atherothrombotic debris into the distal ICA where it embolised and occluded the MCA mainstem. For obvious reasons, this stroke could not have been prevented by completion angioscopy. The second patient suffered an on-table carotid dissection. He had normal TCD flows as the drapes were coming off, but he then took some time to recover from anaesthesia. TCD was reapplied and flow patterns were now consistent Figure 5. Angioscopy showing large intimal flap within the distal internal carotid artery. with an acute carotid occlusion (MCAV similar to that during carotid clamping). This phenomenon occurred in the absence of embolisation during neck closure, which is what one would have expected with a thrombotic occlusion. The neck was reopened and an on-table angiogram revealed an extensive distal dissection, starting about 3 cm above the endarterectomy zone (ie well above the clamp zone). Despite aggressive efforts to repair this surgically, it proved necessary to ligate the ICA at the skull base and he suffered a fatal stroke. It is assumed that the dissection may have been due to an intimal tear caused by the distal Pruitt Inahara shunt balloon (ie could never be prevented by angioscopy). If a similar problem were to occur in the future (flows consistent with ICA occlusion in the absence of preceding embolisation), a diagnosis of distal dissection would be suspected and we would adopt an endovascular treatment strategy. In the final period of study, one patient recovered from anaesthesia with a stroke. This patient had presented with a disabling (Rankin Grade 3) stroke and we were compelled to undertake expedited surgery within seven days of the index symptom because Figure 4. (a) Angioscopy of the proximal CCA region where there is a large thrombus adjacent to the proximal common carotid artery. (b) Once aspirated, the source of the thrombus was seen to be bleeding from a transected vasa vasorum. Reproduced with permission from Naylor et al. J Vasc Surg 2000;32;750e9.

5 R. Sharpe et al. / European Journal of Vascular and Endovascular Surgery 43 (2012) 139e Table 3 Prevalence of 30-day stroke and death following carotid endarterectomy e e e e Total 1995e2010 Intra-operative stroke 1/400 (0.25%) 1/400 (0.25%) 2/400 (0.5%) 1/400 (0.25%) 5/1600 (0.3%) 30-day death 4/400 (1.0%) 2/400 (0.5%) 5/400 (1.25%) 0/400 (0.0%) 11/1600 (0.7%) 30-day death or stroke 8/400 (2.0%) 10/400 (2.5%) 12/400 (3.0%) 3/400 (0.75%) 33/1600 (2.1%) he suffered recurrent TIAs (despite best medical therapy) and he had TCD evidence of ongoing embolisation. At operation, his ICA was too narrow to insert a shunt. His mean MCAV was consistently <15 cm/s throughout the procedure, completion angioscopy was normal and we assume that he suffered a haemodynamic extension to the ischaemic penumbra around his pre-existing cerebral infarct. He went on to make a reasonably good recovery following surgery and was Rankin Grade 2 at 30 days. Table 4 also lists the causes of 28 other strokes/deaths which occurred following a full recovery from anaesthesia (ie postoperative events). Twelve (43%) followed intracranial haemorrhage or the hyperperfusion syndrome (ie they could not have been prevented by angioscopy). Eight post-operative strokes or deaths (29%) were cardiac related and could not have been prevented by angioscopy. These included four deaths after cardiac arrest or Table 4 Causes of stroke and death within 30-days of carotid endarterectomy to Intra-operative events On-table ICA thrombosis MCA occlusion after shunt inserted Haemodynamic, shunt not used a MCA occlusion after flow restoration b Acute on-table ICA dissection 1 nondisabling stroke to disabling stroke b to fatal ischaemic stroke 1 fatal ischaemic stroke Post-operative deaths Intracranial haemorrhage Acute myocardial 1 infarction Acute cardiac 1 failure Cardiac arrest 1 2 Post-operative non-fatal strokes Intracranial haemorrhage Watershed stroke 1 1 after cardiac event MCA embolism Cardiac embolism 1 Hyperperfusion syndrome 2 1 Contralateral ICA 1 embolism Unknown to nondisabling stroke a a Patient had suffered a disabling stroke within 7 days of undergoing surgery and suffered recurrent TIAs. The ICA was too narrow to permit shunt insertion. MCA flows were <15 cm/s throughout the endarterectomy period. b No angioscope available in this case. myocardial infarction; 2 deaths following acute cardiac failure and two haemodynamic strokes following severe acute cardiac failure. The remaining four strokes in the post-operative period were either of unknown aetiology (n ¼ 3) or affected the contralateral hemisphere (n ¼ 1). Four strokes were consistent with post-operative MCA embolisation. These four patients all underwent a normal completion angioscopy and each developed increasing rates of embolisation in the post-operative period prior to onset of their neurological deficit. Discussion The role of intra-operative monitoring during CEA remains an unresolved controversy. There are, of course, ardent supporters of this practice, but if one looks at very large scale studies (rather than small, single centre reports), the evidence would suggest that intra-operative imaging confers no meaningful benefit in terms of reducing peri-operative stroke or death. 22,23 However, while the conclusions from these studies might constitute a convenient reason to dismiss any role for monitoring and/or quality control assessment during CEA, this would actually be an inappropriate interpretation of the data. If one looks at death/stroke rates after CEA, it is highly unlikely that an indiscriminately applied policy of intra-operative imaging or monitoring is going to radically improve results. However, there are important methodological limitations in the way that the role of intra-operative imaging has been evaluated in studies like EVEREST 23 (1305 CEAs) and the New York State Carotid Artery Surgery (NYCAS) study 22 (5988 CEAs) and these have important implications regarding how the current study should be interpreted. The first of these limitations is a failure to define what is actually meant by intra-operative monitoring ; specifically, to differentiate between those modalities that monitor, as opposed to those which fulfil a quality control role. The second important limitation is a failure to discriminate between intra- and postoperative strokes. These two entities have completely different aetiologies and it is illogical to assume that any single monitoring modality could prevent many or all of the strokes or deaths from happening. Monitoring is not the same as quality control (Tables 5 and 6). Accordingly, unless one specifically targets the monitoring/quality Table 5 Intra-operative monitoring during carotid endarterectomy. Detection of Detection of reduced thromboembolism cerebral perfusion TCD Reduced stump pressure Reduced back flow Near infrared spectroscopy Jugular venous O 2 saturation Xenon CBF measurement TCD Detection of loss of electrical activity EEG SSEP Detection of neuronal injury Locoregional anaesthesia

6 144 R. Sharpe et al. / European Journal of Vascular and Endovascular Surgery 43 (2012) 139e145 Table 6 Quality control assessment during carotid endarterectomy. Embolisation Shunt malfunction Luminal thrombus Intimal flap Distal stenosis TCD TCD Angioscopy Angioscopy CW Doppler Angiography Angiography Angiography Duplex Duplex Duplex CW Doppler control strategy to the question being asked, it is unreasonable to expect any strokes to be prevented. For example; CEA performed under locoregional anaesthesia will prevent strokes due to haemodynamic failure during carotid clamping, but will not prevent strokes due to embolisation of retained thrombus following restoration of flow. That would require some form of quality control assessment (eg; angioscopy, Duplex ultrasound or angiography). Similarly, it is unreasonable to expect that intra-operative monitoring using completion assessment with Duplex ultrasound will prevent haemorrhagic strokes or strokes due to the hyperperfusion syndrome in the post-operative period. In a series of themed projects, this unit has shown that a targeted strategy of monitoring and quality control assessment was associated with a 60% decline in the 30-day risks of death/stroke after CEA. 11 One of the earliest findings was that intra-operative stroke (which historically had a higher prevalence than postoperative stroke 24 ) was associated with inadvertent technical error 25 and was almost completely abolished using a combination of TCD and completion angioscopy. 4,5,11 In our experience, the most common cause of intra-operative stroke was embolisation of retained luminal thrombus following restoration of flow, 5 with the source of these thrombi being bleeding from the transected vasa vasorum onto the endarterectomy zone 5 (Fig. 4). In this respect, the introduction of angioscopy to identify and remove these thrombi was associated with an immediate reduction in the risk of intraoperative stroke from 4% to 1%. 6 These thrombi can be quite large and are generally resistant to blind irrigation with heparinised saline. Some colleagues retain concerns that the actual performance of angioscopy may unnecessarily prolong the second clamp period, thus predisposing towards a haemodynamic stroke. This study suggests that such concerns are groundless. Angioscopy takes, on average, about 45 s to perform and only one patient in this series suffered a haemodynamic intra-operative stroke and this was in a patient in whom a shunt could not be inserted. However, the identification and correction of technical error had little or no impact on post-operative stroke and death, especially in those patients suffering strokes due to post-operative carotid thromboembolism. 6 Each of the latter patients was re-explored (after having undergone normal intra-operative angioscopies) and was found to have adherent platelet rich thrombus in the absence of technical error. 6 Subsequent research has shown that patients destined to suffer strokes due to early post-operative thromboembolism have increasing rates of embolisation prior to onset of symptoms 12e19 and that the platelets of high rate embolisers are more sensitive to ADP. 7 In a subsequent research programme, these strokes were prevented by the administration of intravenous Dextran 20 or, more recently, the administration of 75 mg Clopidogrel the night before surgery in addition to regular Aspirin therapy. 8,9 These findings, therefore, highlight the fallacy of assuming that intra-operative monitoring could prevent these strokes, when what they actually required was a novel dual antiplatelet therapy regime starting the night before. Accordingly, the aims of the current study were to see whether; (i) the prevalence of technical error detected using completion angioscopy had diminished with increasing experience (ie was it still really necessary to undertake angioscopy) and (ii) was the extremely low rate of intra-operative stroke maintained through the ensuing 15 year period after the initial pilot and audit studies? Note that the aim of this study was not to see whether completion angioscopy abolished all strokes associated with CEA. As can be seen from Table 2, the vast majority of patients underwent a normal angioscopy. However, 7% had thrombus aspirated prior to restoration of flow. This figure includes both small and large thrombi and it is accepted that an indeterminate proportion of small thrombi will not cause any clinical problem. However, it was very easy to aspirate these retained thrombi (regardless of size) and it is unlikely that any surgeon would wish to take the chance of ignoring the much larger thrombi shown in Figs. 3 and 4. On a practical point, there was never any need to take down the anastomosis in order to aspirate thrombus. They were aspirated and the endarterectomy zone rechecked with the angioscope. Interestingly, the prevalence of detecting luminal thrombus prior to restoration of flow did not decline significantly during the 15-year period of study. This shows that it is still possible for highly experienced surgeons to be associated with inadvertent technical error. Conversely, the prevalence of intimal flaps did decline with increasing experience. In the first 400 patients, large intimal flaps were repaired in almost 5% of patients. By 2010, large intimal flaps were only identified and repaired in 0.8% of patients (p ¼ ). Accordingly, the first conclusion from this study was that while the prevalence of intimal flaps had diminished with time, the same was not true regarding the detection of retained luminal thrombus. The second aim was to determine whether the low rate of intraoperative stroke observed in the pilot study was maintained. Prior to the pilot study (which was undertaken in 1992e1993), the prevalence of intra-operative stroke in our Unit was 4%. 4 In the pilot study (100 patients 4 ) and also in a subsequent audit study (252 patients 5 ) where all CEAs were performed with intra-operative TCD and completion angioscopy, the prevalence of intra-operative stroke fell to almost zero. This extremely low risk of intra-operative stroke has now been maintained for the last 15 years (Table 3) and was 0.3% for the 1600 patients under study. Out of the five intra-operative strokes that did occur, three could definitely not be attributed to any failure of our monitoring/completion angioscopy programme (acute MCA occlusion immediately following shunt insertion, haemodynamic stroke because a shunt could not be inserted, distal ICA dissection). One patient suffered several episodes of on-table thrombosis (despite undergoing a normal angioscopy) and had no evidence of technical error on angiography. There was, however, one definite failure in our programme. The angioscope was not available for use in 117 patients in this series and one suffered a disabling stroke due to an acute MCA occlusion immediately following restoration of flow. This patient almost certainly suffered embolisation of retained thrombus from the endarterectomy zone which would otherwise have been identified and removed had the angioscope been available. In conclusion, despite considerable experience, this study has demonstrated that while the prevalence of large intimal flaps has reduced, the prevalence of retained luminal thrombus remained relatively constant throughout the 15-year period of study. However, the policy of completion angioscopy has been associated with an extremely low prevalence of intra-operative stroke; ie achieving exactly what it was designed to do. While this unit has now stopped performing routine post-operative TCD monitoring and selective Dextran therapy (because dual antiplatelet therapy has now virtually abolished post-operative thromboembolic stroke 9 ), completion angioscopy continues. If nothing else, this study has shown that even the most experienced surgeon can be responsible for causing inadvertent technical error.

7 R. Sharpe et al. / European Journal of Vascular and Endovascular Surgery 43 (2012) 139e Conflict of Interest None. References 1 European Carotid Surgery Trialists Collaborative Group. Randomised trial of endarterectomy for recently symptomatic carotid stenosis: final results of the MRC European Carotid Surgery Trial (ECST). Lancet 1998;351:1379e87. 2 Asymptomatic Carotid Surgery Trial Collaborators. The MRC Asymptomatic Carotid Surgery Trial (ACST): carotid endarterectomy prevents disabling and fatal carotid territory strokes. Lancet 2004;363:1491e Naylor AR, Rothwell PM, Bell PRF. Overview of the principal results and secondary analyses from the European and the North American randomised trials of carotid endarterectomy. Eur J Vasc Endovasc Surg 2003;26:115e29. 4 Gaunt ME, Martin PJ, Smith JL, Rimmer T, Cherryman G, Ratliff DA, et al. The clinical relevance of intra-operative embolisation detected by transcranial Doppler monitoring during carotid endarterectomy: a prospective study in 100 patients. Br J Surg 1994;81:1435e9. 5 Lennard N, Smith JL, Gaunt ME, Abbott R, London NJM, Bell PRF, et al. A policy of quality control assessment reduces the risk of intra-operative stroke during carotid endarterectomy. Eur J Vasc Endovasc Surg 1999;17:234e40. 6 Gaunt ME, Smith JL, Martin PJ, Ratliff DA, Bell PRF, Naylor AR. A comparison of quality control methods applied to carotid endarterectomy. Eur J Vasc Endovasc Surg 1996;11:4e11. 7 Hayes PD, Box H, Tull S, Gaunt ME, Bell PRF, Goodall AH, et al. The patients thrombo-embolic response following carotid endarterectomy is related to enhanced platelet sensitivity to ADP. J Vasc Surg 2003;38:1226e31. 8 Payne DA, Jones CI, Hayes PD, Thompson MM, London NJM, Bell PRF, et al. Beneficial effects of Clopidogrel combined with Aspirin in reducing cerebral emboli in patients undergoing carotid endarterectomy. Circulation 2004;109:1476e81. 9 Sharpe RY, Mennis SJ, Nasim A, McCarthy M, Sayers RD, Lonson NJ, et al. Dual antiplatelet therapy prior to carotid endarterectomy reduces post-operative embolisation: post-operative transcranial Doppler monitoring is now unnecessary. Eur J Vasc Endovasc 2010;40:162e7. 10 Halsey JH, McDowell HA, Gelmon S, Morawetz RB. Blood flow velocity in the middle cerebral artery and regional cerebral blood flow during carotid endarterectomy. Stroke 1989;20:53e8. 11 Naylor AR, Hayes PD, Allroggen H, Lennard N, Gaunt ME, Thompson MM, et al. Reducing the risk of carotid surgery: a seven year audit of the role of monitoring and quality control assessment. J Vasc Surg 2000;32:750e9. 12 Gaunt ME, Smith J, Martin PJ, Ratliff DA, Bell PRF, Naylor AR. On-table diagnosis of incipient carotid artery thrombosis during carotid endarterectomy using transcranial Doppler sonography. J Vasc Surg 1994;20:104e7. 13 Gaunt ME, London NJM, Smith J, Martin PJ, Bell PRF, Naylor AR. Early diagnosis of post-operative carotid occlusion using transcranial Doppler ultrasound. J Vasc Surg 1994;20:1004e5. 14 Levi CR, O Malley HM, Fell G, Roberts AK, Hoare MC, Royle JP, et al. Transcranial Doppler detected cerebral embolism following carotid endarterectomy: high microembolic signal loads predict post-operative cerebral ischaemia. Brain 1997;120:621e9. 15 Cantelmo NL, Babikian VL, Samaraweera RN, Gordon JK, Pochay VE, Winter MR, et al. Cerebral microembolism and ischaemia changes associated with carotid endarterectomy. J Vasc Surg 1998;27:1024e Spencer MP. Transcranial Doppler monitoring and causes of stroke from carotid endarterectomy. Stroke 1997;28:685e Laman DM, Wieneke GH, van Dujin H, van Huffelen AC. High embolic rate after carotid endarterectomy is associated with early cerebrovascular symptoms, especially in women. J Vasc Surg 2002;36:278e Abbott AL, Levi CR, Stork JL, Donnan GA, Chambers BR. Timing of clinically significant microembolisation after carotid endarterectomy. Cerbrovasc Dis 2007;23:362e7. 19 Horn J, Naylor AR, Laman M, Chambers B, Stork J, Schroeder TV, et al. Identification of patients at risk for ischaemic complications after carotid endarterectomy with TCD monitoring. Eur J Vasc Endovasc Surg 2005;30:270e4. 20 Lennard N, Smith JL, Abbott R, Evans DH, London NJM, Bell PRF, et al. Prevention of post-operative thrombotic stroke after carotid endarterectomy: the role of transcranial Doppler ultrasound. J Vasc Surg 1997;26:579e Levi CR, Stork JL, Chambers BR, Abbott AL, Cameron HM, Peeters A, et al. Dextran reduces embolic signals after carotid endarterectomy. Ann Neurol 2001;50:544e7. 22 Rockman CB, Halm EA. Intra-operative imaging: does it really improve peri-operative outcomes of carotid endarterectomy? Sem Vasc Surg 2007; 20:236e Zanetti S, Cao P, DeRango P, Giordano G, Parlani G, Lenti M, et al, And collaborators of the EVEREST Study Group. Eur J Vasc Endovasc Surg 1999; 18:52e8. 24 Krul JM, van Gijn J, Ackerstaff RG, Eikelboom BC, Theodorides T, Vermeulen FE. Site and pathogenesis of infarcts associated with carotid endarterectomy. Stroke 1989;20:324e8. 25 Riles TS, Imparato AM, Jacobowitz GR, Lamparello PJ, Giangola G, Adelman MA, et al. The cause of peri-operative stroke after carotid endarterectomy. J Vasc Surg 1994;19:206e14.